Review Reports

Round 1

Reviewer 1 Report

 

The paper provides a new hybrid forecasting methodology for improving the short-term electrical load forecasting accuracy. While the paper provides an interesting methodological approach, there are some critical weaknesses that need to be addressed.

Major Comments

1.       A comparative Table should be provided that identifies the issues addressed by each academic paper reviewed. Through this comparative Table the authors need to clearly depict in bullets, the research gaps identified from the literature review analysis and document how does their methodology address the research gaps on a one-to-one basis.

2.       The modelling methodology is difficult to follow. The authors therefore need to provide a nomenclature Table describing the model parameters employed.

3.       gaps identified from the literature review analysis and document how does their methodology address the research gaps on a one-to-one basis.

4.       The modelling methodology is difficult to follow. The authors therefore need to provide a nomenclature Table describing the model parameters employed.

ology for improving the short-term electrical load forecasting accuracy. While the paper provides an interesting methodological approach, there are some critical weaknesses that need to be addressed.

Major Comments

5.       A comparative Table should be provided that identifies the issues addressed by each academic paper reviewed. Through this comparative Table the authors need to clearly depict in bullets, the research gaps identified from the literature review analysis and document how does their methodology address the research gaps on a one-to-one basis.

6.       The modelling methodology is difficult to follow. The authors therefore need to provide a nomenclature Table describing the model parameters employed.

 

Author Response

Dear reviewer

I have revised the manuscript according to your comments, please refer to the attachment for specific modifications.

Yours sincerely

Dr. Yang

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments

The authors propose a Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Multiverse optimization algorithm (MVO), and the Gated Recurrent Unit (GRU) based on Rectified Adam (RAdam) optimizer to solve the short-term load Forecasting problem. Electric load or demand forecasting is a hot research topic. But this article still has some room for improvement, my comments and suggestions are as follows:

1a. It is suggested to remove abbreviations from title.

1. Why use the the GRU network model based on RAdam optimizer as the backbone network, and whether there is relevant literature or theoretical support. With the development of deep learning technology, many models suitable for time series have also been developed, such as TCN.

2. The Schematic diagram of the MVO and RAdam in Figures 1 and 2 are not standardized.

2b. Figure 4 is at abstract level. Highlight the portion of MVO, GRU, and Radam t in the flow chart shown in Figure 4. It does not convey exact meaning of the developed model.

3. This paper should make a clear contribution. MVO, GRU, and Radam etc. are all commonly used methods. What are the advantages of their joint use? Are they necessary to be used in combination? In the load forecasting framework, are they isolated or mutually reinforcing.

4. How are the model parameters determined?

5. This paper emphasizes the improvement of the accuracy many times. Is there any experiment to clarify its value other than MAPE and Spearman correlation coefficient? I have a doubt whether the model MAPE and Spearman correlation coefficient is a suitable evaluation metric, because it does not affect the final performance of the model. And can you cite some other literature focusing on the accuracy test. Also, I suggest to add metrics speed of model convergence and clarify its importance.

6. It is suggested that the abstract be improved to highlight the value of the proposed method in this paper.

7. Literature review section must be strengthening by including latest papers for example

-https://doi.org/10.3390/en13092244

Author Response

Dear reviewer

I have revised the manuscript according to your comments, please refer to the attachment for specific modifications.

Yours sincerely

Dr. Yang

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper proposes a hybrid CEEMDAN-MVO-GRU model for short-term power load prediction. The authors use CEEMDAN to decompose the original electric load series, solving the problem of randomization of short-term load fluctuations and difficulty in selecting influencing factors. Based on this, the GRU network is used for the prediction of modal components. Meanwhile, the MVO optimization algorithm is adopted to optimize the GRU network parameters for different sequences of prediction problems. Comparative analysis show that the proposed model is superior to its counterparts which has important application value in the power load prediction.

This paper is well prepared, in terms of language, writing, and organization. To improve quality of the manuscript my suggestions are listed below:

(1) The discussion can be improved by comparing the study findings with previous studies.

(2) The conclusion section could be more fleshed out. You can add some indications of power grid adjustment based on your findings.

(3) The meaning of Figure 13.b is not explained. If you don't use this figure, you can delete it.

(4) The citations in this paper, including literature, pictures, and tables, cannot be transferred to the corresponding content.

(5) No spaces are required before “where s(t) is the original signal......”. The same is true: “where rho is...... ” ......

(6) It is recommended that the best results in Table 3 be shown in bold.

(7) The format of the references should be uniform.

(8) Please update some recent work to enhance your literature review. I have searched for some references for your consideration as:

a.     Yang, Yang, et al. "An effective dimensionality reduction approach for short-term load forecasting." Electric Power Systems Research 210 (2022): 108150.

b.     Zhao, Shangrui, et al. "An asymmetric bisquare regression for mixed cyberattack-resilient load forecasting." Expert Systems with Applications 210 (2022): 118467.

c.     Yang, Yang, et al. "An integrated federated learning algorithm for short-term load forecasting." Electric Power Systems Research 214 (2023): 108830.

d.     Yang, Yang, et al. "A novel decompose-cluster-feedback algorithm for load forecasting with hierarchical structure." International Journal of Electrical Power & Energy Systems 142 (2022): 108249.

 

Some typing errors:

Page 11: “IFM is shown” instead of “IMF”

Equation 16: “n” instead of “N”

Author Response

Dear reviewer

I have revised the manuscript according to your comments, please refer to the attachment for specific modifications.

Yours sincerely

Dr. Yang

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

I have no more  comment.